Aalborg Universitet Radio Resource Management for 5G Small Cells in Unpaired Spectrum With Focus on Dynamic TDD and Full Duplex Technology Gatnau, Marta

(1)

Aalborg Universitet

Radio Resource Management for 5G Small Cells in Unpaired Spectrum With Focus on Dynamic TDD and Full Duplex Technology

Gatnau, Marta

DOI (link to publication from Publisher):

10.5278/vbn.phd.engsci.00147

Publication date:

2016

Document Version

Publisher's PDF, also known as Version of record Link to publication from Aalborg University

Citation for published version (APA):

Gatnau, M. (2016). Radio Resource Management for 5G Small Cells in Unpaired Spectrum: With Focus on Dynamic TDD and Full Duplex Technology. Aalborg Universitetsforlag. Ph.d.-serien for Det Teknisk- Naturvidenskabelige Fakultet, Aalborg Universitet https://doi.org/10.5278/vbn.phd.engsci.00147

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

- Users may download and print one copy of any publication from the public portal for the purpose of private study or research.

- You may not further distribute the material or use it for any profit-making activity or commercial gain - You may freely distribute the URL identifying the publication in the public portal -

Take down policy

If you believe that this document breaches copyright please contact us at vbn@aub.aau.dk providing details, and we will remove access to the work immediately and investigate your claim.

(2)

(3)

RADIO RESOURCE

MANAGEMENT FOR 5G SMALL CELLS IN UNPAIRED SPECTRUM

– WITH FOCUS ON DYNAMIC TDD AND FULL DUPLEX TECHNOLOGY

MARTA GATNAU SARRETBY DISSERTATION SUBMITTED 2016

RADIO RESOURCE MANAGEMENT FOR 5G SMALL CELLS IN UNPAIRED SPECTRUMMARTA GATNAU SARRET

(4)

(5)

Radio Resource

Management for 5G Small Cells in Unpaired Spectrum

- With Focus on Dynamic TDD and Full Duplex Technology

Ph.D. Dissertation

Marta Gatnau Sarret

Aalborg University Department of Electronic Systems

Fredrik Bajers Vej 7 DK – 9220 Aalborg

(6)

Dissertation submitted: August, 2016

PhD supervisor: Prof. Preben Mogensen

Aalborg University

Assistant PhD supervisors: Assoc. Prof. Gilberto Berardinelli

Aalborg University

Post. Doc. Nurul Huda Mahmood

Aalborg University

PhD committee: Associate Professor Tatiana K. Madsen (chairman)

Aalborg University

Professor Mikko Valkama

Tampere University of Technology

Head of Network Technology Lab dr. Peter Karlsson

Sony Mobile

PhD Series: Faculty of Engineering and Science, Aalborg University

ISSN (online): 2246-1248

ISBN (online): 978-87-7112-779-9

Published by:

Aalborg University Press Skjernvej 4A, 2nd floor DK – 9220 Aalborg Ø Phone: +45 99407140 aauf@forlag.aau.dk forlag.aau.dk

Printed in Denmark by Rosendahls, 2016

(7)

Abstract

Wireless communication is driving a networked society, where data is ex- changed anytime, everywhere, between everyone and everything. The inclu- sion of smart devices and the explosion of new applications requiring mobile connectivity entails ambitious requirements for the next 5^th Generation (5G) system: data rates up to 10 Gbps, latency as low as 1 millisecond and support for new use cases such as Device-to-Device (D2D) Communication.

Several solutions may be considered to achieve higher system capacity:

using larger spectrum bands, increasing the number of cells and improving the spectral efficiency. The first approach has limited potential, since the spectrum is scarce and expensive at conventional frequency bands. The strategy of increasing the number of radio cells and reducing their size has proven to be promising. However, small cells serve a small set of users, creating sudden traffic imbalances between uplink and downlink directions. Thus, to optimally accommodate the instantaneous traffic demands, dynamic Time Division Duplex (TDD) is the most appropriate transmission mode. Improv- ing the spectral efficiency is usually achieved by adding more antennas in a device. However, this approach may bring constraints in terms of space and cost. On the other hand, Full Duplex (FD) has been positioned as a potential technology for 5G by allowing a node to transmit and receive simultaneously in the same frequency band, thus theoretically improving the spectral efficiency and the system capacity.

Nevertheless, these approaches have a common drawback: increased and unpredictable interference. Denser networks lead to a larger inter-cell interference (ICI) and FD doubles the amount of interfering streams. Therefore, mechanisms to deal with the interference are essential for the 5G design.

The focus of this Ph.D study is on Radio Resource Management (RRM) for 5G small cells in unpaired spectrum, targeting dynamic TDD and FD nodes.

In the first part of the Ph.D. thesis, link adaptation and recovery mechanisms are studied to overcome the drawbacks caused by dynamic TDD. System level simulations show that recovery mechanisms, specially at lower layers, are needed to meet the 5G targets. Moreover, results also show that interference

(8)

cancellation receivers highly alleviate the described interference challenges.

In the second part of this work, an RRM framework supporting FD is designed and studied. The joint performance of dynamic TDD and FD is evaluated and compared in terms of throughput gain and delay reduction.

Two cases of FD are studied: when both the Base Station (BS) and the User Equipment (UE) are FD capable, and when only the BS can exploit FD. In- tensive and detailed system level simulations are carried out, showing that in realistic environments the theoretical FD gain is significantly reduced, mainly due to the traffic profile and the increased ICI.

Finally, motivated by the strict 5G latency target, direct D2D discovery, i.e., the node awareness procedure prior to the communication phase, is studied.

FD is considered as an attractive solution to achieve fast discovery and meet the latency target since a node can continuously receive while still broadcast- ing messages to its neighbors. The study shows that, in order to meet the latency target and get benefits from FD, interference cancellation receivers are a must. In that case, FD can reduce the latency significantly and reach the strict 5G D2D discovery target.

(9)

Resumé

Trådløs kommunikation stimulerer et netværkssamfund, hvor der konstant udveksles data alle vegne, imellem alle mennesker og alle ting. Inkluderingen af smarte apparater, og eksplosionen af nye applikationer, som kræver mobil- forbindelse, medfører ambitiøse krav til det næste 5. generationssystem: Data hastigheder på op til 10 Gbps, forsinkelse ned til 1 millisekund, og support i form af ”Device-2-Device” (D2D) kommunikation til nye brugsmønstre.

Man kan overveje forskellige muligheder for at opnå højere kapacitet på systemet: Brug af større spektrum bånd, forøge antallet af celler, og forbedre spektre-effektiviteten. Det første tiltag giver et begrænset udbytte, da spek- trummet er begrænset og dyrt i de konventionelle frekvensbånd. Det har vist sig at være mere lovende at vælge en strategi, hvor man forøger antallet af radioceller, og formindsker deres størrelse. Men små celler betjener en lille gruppe af brugere, hvilket skaber pludselige trafikale ubalancer imellem uplink og downlink retninger. Så for at imødekomme de øjeblikkelige be- hov for trafik bedst muligt, er ”Time Division Duplex” (TDD) den bedste måde at overføre på. Muligheden for at forbedre spektre-effektiviteten opnås normalt ved at tilføje flere antenner på apparatet, men denne tilgang med- fører begrænsninger i form af volume og pris. På den anden side anses "full duplex" (FD) for at være en mulig teknologi indenfor 5G, ved at tillade at et adgangspunkt sender og modtager samtidigt i det samme frekvensbånd, således at spektre-effektiviteten og systemkapaciteten i teorien forbedres.

Disse tiltag har ikke desto mindre en ulempe: Forøget og uforudsigelig interferens. Tættere netværk fører til en større interferens imellem cellerne (ICI), og FD fordobler antallet af interferens strømninger. Derfor er det vigtigt for designet af 5G netværket, at finde mekanismer til at håndtere interferens.

Fokus i dette PhD studie er lagt på ”Radio Ressource Management” (RRM) til små 5G celler i ”unpaired” spektrum, for at opnå dynamisk TDD og FD. I den første del af PhD afhandlingen har man undersøgt tilpasning af forbindelse og gendannelsesmekanismer, for at imødekomme ulemperne forårsaget af dynamisk TDD. Simuleringer på systemlag viser, at gendannelsesmekanismer – specielt på lavere lag – er nødvendige for at imødekomme 5G målene.

(10)

Desuden viser resultater også at interferens-elliminerende modtagere, som skal afbryde/modstå interferens, i høj grad dæmper de beskrevne udfor- dringer med interferens.

I den anden del af denne afhandling har man designet og undersøgt en RRM struktur, som understøtter FD. Man har evalueret og sammenlignet den samlede performance af dynamisk TDD og FD i forhold til forøgelse af data hastigheder og reduktion af forsinkelse. Der er undersøgt 2 scenarier af FD:

Når både basestationen (BS) og brugerudstyret (UE) har FD – og når kun BS kan udnytte FD. Når der udføres intensive og detaljerede simuleringer på systemniveau viser det, at den teoretiske FD forøgelse reduceres markant i realistiske omgivelser, mest på grund af trafik-profilen og den forøgede ICI.

Til sidst undersøges proceduren for ”adgangspunkt-bevidstheden” forud for kommunikationsfasen, motiveret af det strenge mål for forsinkelse på 5G, og det direkte D2D forskningsresultat. FD anses for at være en attraktiv løsning til at opnå et hurtigt forskningsresultat, og imødekomme målet for forsinkelse, da et adgangspunkt hele tiden kan lytte, alt imens det sender beskeder til sine naboer. Undersøgelsen viser, at for at imødekomme målet for forsinkelse, og for at kunne udnytte fordelene ved FD, skal man bruge modtagere som afviser interferens. I så fald kan FD reducere forsinkelsen betydeligt, og man kan opnå det strenge forskningsresultat for 5G D2D.

Translated by Dorthe Sparre, Aalborg University, Denmark.

(11)

List of Abbreviations

3G 3^th Generation

3GPP 3rd Generation Partnership Project 4G 4^th Generation

5G 5^th Generation

5GPPP 5G Infrastructure Public Private Partnership AM Acknowledged Mode

AP Access Point

ARQ Automatic Repeat Request BLER Block Error Rate

BS Base Station CP Cyclic Prefix

CQI Channel Quality Information D2D Device-to-Device

DL Downlink

DMRS Demodulation Reference Symbol eMBB enhanced mobile broadband FD Full Duplex

FEC Forward Error Correction FTP File Transfer Protocol

(18)

List of Abbreviations

GP Guard Period

GPS Global Positioning System

HARQ Hybrid Automatic Repeat Request HD Half Duplex

HSPA High Speed Packet Access ICI inter-cell interference

IRC Interference Rejection Combining KPI Key Performance Indicator LTE Long Term Evolution MAC Medium Access Control

MCS Modulation and Coding Scheme MIMO Multiple Input Multiple Output MMSE Minimum Mean Square Error

mMTC massive machine type of communication MRC Maximum Ratio Combining

OFDM Orthogonal Frequency Division Multiplexing OLLA Outer Loop Link Adaptation

OSI Open Systems Interconnection PAPR Peak-to-Average Power Ratio PHY Physical

PRB Physical Resource Block

QAM Quadrature Amplitude Modulation RAT Radio Access Technology

RLC Radio Link Control

RRM Radio Resource Management RTT Round Trip Time

SG scheduling grant

(19)

SI Self-Interference

SINR Signal to Noise plus Interference Ratio SR scheduling request

TCP Transmission Control Protocol TDD Time Division Duplex TTI Transmission Time Interval UDP User Datagram Protocol UE User Equipment

UL Uplink

UM Unacknowledged Mode

URLLC ultra-reliable low latency communication V2X Vehicle-to-Anything

(20)

(21)

Thesis Details

Thesis Title: Radio Resource Management Solutions for Unpaired 5G Small Cells - With Focus on Dynamic TDD and Full Du- plex Technology

Ph.D. Student: Marta Gatnau Sarret

Supervisors: Prof. Preben Mogensen, Aalborg University

Assoc. Prof. Gilberto Berardinelli, Aalborg University Post. Doc. Nurul Huda Mahmood, Aalborg University The work conducted in this thesis is the result of three years of research at the Wireless Communication Networks section, Department of Electronic Systems, Aalborg University, Denmark, in close cooperation with Nokia – Bell Labs. The work presented in this thesis was carried out in parallel with the mandatory courses and teaching/working obligations needed to obtain the Ph.D. degree.

The main body of this thesis consist of the following papers:

[A] Marta Gatnau Sarret, Davide Catania, Frank Frederiksen, Andrea Fabio Cattoni, Gilberto Berardinelli, Preben Mogensen, "Improving Link Ro- bustness in 5G Ultra-Dense Small Cells by Hybrid ARQ",IEEE 11th In- ternational Symposium on Wireless Communications Systems (ISWCS), 2014.

[B] Marta Gatnau Sarret, Davide Catania, Frank Frederiksen, Andrea Fabio Cattoni, Gilberto Berardinelli, Preben Mogensen, "Dynamic Outer Loop Link Adaptation for the 5G Centimeter-Wave Concept",IEEE 21th Eu- ropean Wireless Conference (EW), 2015.

[C] Marta Gatnau Sarret, Gilberto Berardinelli, Nurul H. Mahmood, Marko Fleischer, Preben Mogensen, Helmut Heinz, "Analyzing the Potential of Full Duplex in 5G Ultra-Dense Small Cell Networks"EURASIP Journal on Wireless Communications and Networking - Special issue: Full-Duplex

(22)

Thesis Details

Radio: Theory, Design, and Applications.Submitted.

[D] Marta Gatnau Sarret, Gilberto Berardinelli, Nurul H. Mahmood, Preben Mogensen, "Can Full Duplex reduce the discovery time in D2D Com- munication?"IEEE 13th International Symposium on Wireless Communica- tions Systems (ISWCS), 2016.

[E] Marta Gatnau Sarret, Gilberto Berardinelli, Nurul H. Mahmood, Preben Mogensen, "Providing Fast Discovery in D2D Communication with Full Duplex Technology"Springer 9th International Workshop on Multiple Ac- cess Communications (MACOM), 2016.Submitted.

In addition to the main papers, the following invention disclosures were submitted in Nokia:

• Marta Gatnau Sarret, Frank Frederiksen, "Channel Quality Indicator (CQI) reporting procedure for 5G systems".

• Marta Gatnau Sarret, Frank Frederiksen, "Interference Cancelling Aware (ICA) Channel State Indicator (CSI) reporting for 5G systems".

• Marta Gatnau Sarret, Frank Frederiksen, "Separation between protected and non-protected data for 5G systems".

• Marta Gatnau Sarret, Beatriz Soret, "Increasing the reliability by inter-cell signaling in full duplex systems".

• Marta Gatnau Sarret, Beatriz Soret, Gilberto Berardinelli, Nurul Huda Mah- mood, Frank Frederiksen, "Fast discovery for autonomous devices".

This thesis has been submitted for assessment in partial fulfillment of the PhD degree. The thesis is based on the submitted or published scientific papers which are listed above. Parts of the papers are used directly or indirectly in the extended summary of the thesis. As part of the assessment, co-author statements have been made available to the assessment committee and are also available at the Faculty.

(23)

Preface

The research project was financed by Nokia Bell Labs and has been completed under the supervision of Professor Preben Mogensen (Aalborg Univer- sity and Nokia Bell Labs), Associate Professor Gilberto Berardinelli (Aalborg University) and Post. Doc. Nurul Huda Mahmood (Aalborg University).

First of all, I would like to thank my supervisors, for their patience and for supporting and guiding me during this journey. I have learned something different from each of you, which helped me to grow both personally and as an engineer.

I would also like to express my gratitude to the persons who accepted to be part of my Ph.D. assessment committee, Doctor Peter Karlsson, Professor Mikko Valkamma and Associate Professor Tatiana Kozlova Madsen, for taking the time to read and assess this work.

Thanks to my parents, my brother and my sister-in-law, for always be- lieving in me and making me feel close to you, even though we were at 2.500 kilometers. Your visits and your support was what kept me running.

Thanks to my boyfriend, for his patience and positivism. Your company and your ability to make me laugh makes everything easier.

Last but not least, I would like to thank all my colleagues and friends at Aalborg University and Nokia Bell Labs. I have really enjoyed all these years working with you, chating, laughing, having so much fun in Summer and Christmas parties. The memories that I have will remain always with me.

Finally, special thanks my girls, Laura, Isa and Carmen, for being amazing friends and teach me so many things, not engineering related.

Marta Gatnau Sarret Aalborg University August 31, 2016

(24)

Preface

(25)

Part I

Introduction

(26)

(27)

Background and Thesis Overview

1 Mobile Traffic Growth

Over the last few years, a remarkable growth of data-enabled devices has been observed, which has led to an increase of the mobile data traffic. The amount of data carried by mobile networks moved from 10 gigabytes per month in 2010 to 3.7 exabytes per month in 2015 [1]. This outstanding growth has been caused by cellular network advances, moving from the 3^th Gen- eration (3G) to the 4^th Generation (4G) of mobile systems; by the massive increase of smart devices; and by the emergence of new type of applications, such as social networks, whose number of users grows notably day by day. Only from 2014 to 2015, mobile data traffic grew 74%, and even though smart devices with minimum 3G connectivity only represented 36% of the total number of mobile devices and connections, they represented 89% of the mobile data traffic [1].

Such numbers are continuously growing, and the prediction is that global mobile data traffic will increase approximately by eight times between 2015 and 2020, as shown in Figure 1.1. Smart devices are forecasted to generate 98% of the mobile data traffic by 2020 [1], indicating that devices using wireless data connection are becoming more popular year by year.

To increase the network capacity and accommodate the future traffic demands, the strategies that must be considered, shown in Figure 1.2, are:

• Using larger frequency bands

• Enhancing the spectral efficiency

• Increasing the cell density

(28)

Background and Thesis Overview

2015 2016 2017 2018 2019 2020 0

5 10 15 20 25 30 35

Exabytes per month

3.7 EB 6.2 EB

9.9 EB 14.9 EB

21.7 EB 30.6 EB

Fig. 1.1:Exabytes per Month of Mobile Data Traffic - Source: Cisco VNI 2016 [1].

Fig. 1.2:Possible strategies to accommodate the future traffic demands.

The frequencies ranging from 3 to 30 GHz, also known as centimeter- wavefrequencies, have the most attractive propagation characteristics. How- ever, using this frequency range has limited potential because the spectrum is scarce and expensive, making it insufficient to deal with the future traffic growth [2]. On the other hand, higher frequency bands ranging from 30 up to 300 GHz, also referred as millimeter-wave frequencies, have been considered as an option to increase the network capacity [3, 4]. Research shows that there is potential for using such bands. However, given the different propagation conditions compared to the traditionalcentimeter-wavefrequencies, large antenna arrays and beamforming techniques are required to overcome the pathloss. Orthogonal Frequency Division Multiplexing (OFDM) might not be the most suitable modulation formillimeter-wave. This modulation is very appropriate for bandwidth limited systems to multiplex users in frequency.

However, since a large amount of spectrum is available in themillimeter-wave frequencies, users can be multiplexed in time. There are already some pro- posals on a modulation for these high frequencies [5], as well as discussions on channel modeling and beamforming design [3].

To enhance the spectral efficiency, the most common and used approach is

(29)

2. Introduction to 5G

to exploit Multiple Input Multiple Output (MIMO) antenna technology. This technique allows to overcome the effects of the signal multipath and fading in systems with limited bandwidth that require high throughput. The main drawback of such technology is that, as the number of device antennas grows, the constraints in terms of space and cost increase. Thus, it is unlikely that future devices will be equipped with a large number of antennas. Another option that could be considered is the emerging Full Duplex (FD) technology. It allows a device to transmit and receive simultaneously in the same frequency band, thus, theoretically, doubling the throughput compared to conventional Half Duplex (HD) systems. Currently, the main concern about FD technology is the cancellation of the Self-Interference (SI) power, which refers to the interference generated by the transmitted signal at the receiver end of the same node. The SI power must be highly attenuated in order to have an operational FD transceiver. Nowadays, the advances on transceiver design show that approximately 110 dB of SI cancellation are possible to be achieved [6], according to a certain bandwidth and power constraints, thus enabling FD as a potential technology to provide better spectral efficiency.

Increasing the cell density has been positioned as the other key solution to deal with the expected mobile traffic growth. It is achieved by deploying a large number of low-power base stations in scenarios with high traffic density, while traditional macro cell base stations provide basic service coverage.

According to [2, 7], a massive deployment of small cells is fundamental for moving towards the desired increase in network capacity. Nowadays, the number of small cell deployments already exceeds the number of macro-cell base stations [8], and according to [9], most of the traffic is generated in- doors. This statements reinforces the necessity of using indoor small cells with short ranges, even though the deployment of outdoor small cells has already been started by some operators to improve the coverage performance of the macro-cells [8].

2 Introduction to 5G

Current mobile systems may not be able to accommodate the expected traffic demands by 2020. Consequently, both the industry and the academia have started developing a concept for the future 5^th Generation (5G) Radio Ac- cess Technology (RAT). This concept is debated as either an evolution or a revolution, i.e., a 5G system conceived as an improvement of the current 4G system or a disruptive new RAT, respectively. Ericsson believes that 5G will be an improvement of the current 4G system [10]. On the other hand, Huawei is confident that 5G will emerge as a new RAT [11]. The European Project 5G Infrastructure Public Private Partnership (5GPPP), composed of members

(30)

from both academia and industry, also supports the idea of 5G being a new disruptive system [12]. Independently of the concept, the targets of the system are the same: peak data rates of 10 Gbps, extremely low latencies of 1 millisecond, coverage everywhere, and support for ultra-high reliability and ultra-low latency applications, such as Vehicle-to-Anything (V2X).

The 5G standardization process has already been started, given the fact that the system should be ready for deployment by 2020. Figure 1.3 shows the anticipated time line from research to standardization and commercial- ization. According to the defined time schedule, there will be a first trial of the basic 5G functionality by 2017 and a second one by 2018.

Fig. 1.3:5G time line from research to standards - Source: Nokia 2015.

This dissertation is part of a large project in Nokia Bell Labs. The position of Nokia Bell Labs is to design 5G as a new disruptive RAT, since the limits of 4G technologies are being rapidly approached. The 5G vision of the company is depicted in Figure 1.4. The figure shows the use cases that 5G should support and the corresponding system targets. The first use case is enhanced mobile broadband (eMBB). The target is to improve the system capacity to provide peak data rates of 10 Gbps, whereas getting 100 Mbps must be possible whenever and wherever. The second use case is the ultra-reliable low latency communication (URLLC). Very strict requirements are envisioned, such as latencies below 1 millisecond and 99.999% reliability [13]. Finally, to support the third use case, namely massive machine type of communication (mMTC), the system should be optimized to handle a large amount of low cost devices with rigorous energy consumption requirements. It is important to notice that those requirements cannot be fulfilled simultaneously due to fundamental theoretical limits [14], but 5G aims at a flexible design which is able to cope with these requirements on a link basis.

In terms of spectrum band, there might be frequencies that are more suitable for certain type of applications than others. According to [15], frequencies below 6 GHz can provide support for the three use cases, eMBB, URLLC and mMTC. In the range between 3 and 40 GHz, the main focus is on eMBB, whereas URLLC could still be considered. Then, in frequencies above 40 GHz, only the eMBB use case can be addressed.

For the eMBB use case, where the main objective is to maximize the system capacity, an heterogeneous deployment is envisioned [16]. It refers to

(31)

2. Introduction to 5G

Fig. 1.4:5G Nokia Vision with the corresponding targets - Source: Nokia 2015.

a combination of macro sites, outdoor micro cells and indoor small cells, as shown in Figure 1.5, which also indicates the ratio of cell density increase among the different radio cells. The target of the macro sites is to provide basic service coverage, to be able to ensure voice and data services. Outdoor micro cells play an important role to achieve high data rate everywhere, e.g., 10 Mbps with 90% coverage. Indoor small cells have the largest amount of traffic share [9], but the signal propagation is more sensitive to attenuations due to the floor and walls, and the transmit power is lowered due to health regulations. Since the indoor coverage may be limited, a large number of cells should be deployed to efficiently accommodate the wireless traffic.

Fig. 1.5:Heterogeneous network, with deployment ratios according to [16].

(32)

Figure 1.4 shows that the 5G scope is very broad. This dissertation addresses two topics: the eMBB, in particular the performance of 5G small cells in unpaired spectrum; and the URLLC, more precisely how to provide fast discovery in Device-to-Device (D2D) communication. Next section presents the literature review on these topics.

3 Literature review

The general state of the art of the research areas of the project is divided in three sections: the performance of dynamic Time Division Duplex (TDD) in small cells, the potential of full duplex technology in small cells and how to accelerate the discovery procedure of D2D communication.

Small Cells and Dynamic TDD

The potential of small cells, i.e., of increasing the cell density for boosting the network capacity has already been proven [2, 7, 8]. As the cell size shrinks, the number of users served by the Access Point (AP) is reduced. Therefore, since the level of flow aggregation is low, an unpredictable traffic burtiness is expected. To optimally accommodate such traffic demands, the most appropriate solution would be to provide dynamism in terms of transmission direction, independently in each cell. This technique is commonly known as dynamic TDD.

The feasibility of dynamic TDD has been addressed in the literature.

In 3rd Generation Partnership Project (3GPP) Release 12, 7 configurations with different Uplink (UL)/Downlink (DL) asymmetries are defined for Long Term Evolution (LTE), named LTE-TDD [17]. The proposed configurations and their corresponding switching times are shown in Figure 1.6. In [18], the evaluation of LTE-TDD with a switching point periodicity of 10 milliseconds is evaluated, showing that providing dynamism in terms of transmission direction gives higher gains than using a static allocation. In addition, the authors conclude that the system performance improves as the switching point periodicity reduces. On the other hand, the study also indicates that interference mitigation techniques are required, since dynamic TDD generates unpredictable interference.

In [19], an uncoordinated and greedy dynamic TDD scheme is evaluated.

The study shows the necessity for interference mitigation and management techniques, given the interference variability caused by dynamic TDD. The authors use a beamsteering strategy to reduce the impact of the interference

(33)

3. Literature review

Fig. 1.6:LTE-TDD configurations defined by 3GPP in release 12. Source: [17].

without inter-cell coordination. Results show that dynamic TDD outperforms an approach with a static allocation of the transmission direction. The work presented in [20] addresses the use of interference cancellation to mitigate the problem introduced by dynamic TDD. Results show that the use of techniques to reduce the interference are required to benefit from dynamic TDD.

In that case, the gains are significant over a non-flexible transmission direction approach.

It is expected that even if advanced techniques to reduce the interference are used, interference cannot be avoided completely. To mitigate the impact of such residual interference, several techniques are investigated, e.g., power control, recovery mechanisms, rank adaptation or interference alignment. In [21], a hybrid solution which combines a dynamic and a static TDD approach with power control is presented. The authors show the benefits of the proposed solution to improve the system capacity. An inter-cell alignment based MIMO transmission scheme is presented in [22]. The proposed strategy combines interference cancellation and interference alignment according to the level of interference. Results show the feasibility of the proposed scheme in improving the system throughput. The well-known Hybrid Automatic Repeat Request (HARQ) recovery mechanism is still considered as an important part for the design of the future RAT. In [23], a novel signaling scheme for HARQ is proposed to deal with the feedback misalignment in the dynamic LTE-TDD system [17].

Other solutions are based on rank adaptation techniques, to improve the system performance by reducing the overall network interference. In [24], a cooperative distributed rank coordination scheme is presented. The goal of the mechanism it to maximize the network utility function instead of the performance of each individual node. The authors assume that cells are coor- dinated following a master-slave architecture, where nodes exchange limited amount of information. A power control strategy to mitigate the interference caused by dynamic TDD in LTE is presented in [25]. The proposed scheme aims at minimizing the AP-to-AP interference to improve the UL throughput performance. Results show that the proposed scheme can improve the UL direction at expenses of a minor degradation in the DL.

(34)

Small Cells and Full Duplex Technology

Given the unpredictable interference caused by dynamic TDD and the advances in the self-interference cancellation techniques [26, 27], FD technology has been positioned as a potential candidate for the future 5G system. For a FD node to be operational, the SI power should be highly attenuated. Recent results show that SI can be reduced approximately by 100 dB [27], which may suffice for considering FD a viable option, according to certain bandwidth and transmit power constraints. Nevertheless, the double throughput gain that FD promises may be compromised by several limitation, namely residual SI, increased interference and traffic asymmetry. The inter-cell interference (ICI) is increased with FD since the amount of interfering streams are doubled. In addition, to exploit FD it is required that data traffic is present on both link directions, i.e., UL and DL.

FD technology is rather expensive and energy consuming, and it could be that for the time being, it would only be implemented in the base stations [28–30]. However, the case where the User Equipments (UEs) are FD capable is also under study [31–33].

In [28], a scheduler that minimizes the UE-to-UE interference is presented, showing that FD can achieve a throughput gain over conventional HD. How- ever, the penetration wall loss assumed in this work mitigates the impact of the inter-cell interference. Note that the penetration wall loss dictates the iso- lation between cells, thus defining the impact of the inter-cell interference on the system performance. The authors in [29] evaluate the performance of FD in ultra-dense small cells. Several user scheduling techniques alongside an optimal power allocation scheme are provided, showing that FD outperforms HD if a certain level of SI cancellation is achieved, e.g., 70 dB. The area spectral efficiency and the coverage in a small cell network are evaluated with HD and FD in [30]. Results show that FD achieves higher area spectral efficiency than HD, but at the expenses of higher outage, i.e., lower coverage. The authors indicate a compromise in the UL and DL transmit power to achieve the optimal performance.

The study presented in [31] evaluate FD in a dense deployment of small cells. The authors conclude that the SI dominates over the inter-cell interference if the SI cancellation capabilities are below 100 dB. In [32], results comparing the performance of MIMO HD and FD in a small cell scenario are discussed. Results show that under strong interference, HD may outperform FD due to MIMO spatial multiplexing gains. Under low interference conditions, FD can improve both the throughput and delay of the system. Finally, the authors in [33] study the achievable bit rate depending on the residual SI power and the inter-cell interference conditions. The work analyzes the Signal to Noise plus Interference Ratio (SINR) region where FD outperforms

(35)

3. Literature review

HD. The authors conclude that in highly interfered scenarios, a strategy that switches between FD and HD provides the optimal system performance.

Note that most of the research work on the performance of full duplex assumes full buffer traffic. Therefore, the impact of the increased interference is studied, but the effect of the traffic asymmetry, the high layers protocols or the jointly repercussion of all the mentioned constraints are not addressed.

Device-to-Device Communication

D2D communication is an important topic in current radio research, since allowing devices to exchange data directly among them helps to offload the infrastructure. However, the discovery phase of such type of communication is not extensively. The discovery procedure refers to the detection of peer devices in the surroundings, which is required to establish a unicast or multicast communication. It can be performed with the aid of the infrastructure or autonomously by the devices. The former option generates additional control signaling and thus increases the overhead. For this reason, in this dissertation, the focus is on an autonomous discovery procedure, since allowing devices to exchange control messages directly can benefit the system in reducing the control overhead and the discovery time. According to the latest specifications for the next generation access technologies, the control plane latency target is set to 10 milliseconds [13].

A discovery mechanism for contention-based networks is proposed in [34]. It defines an initiator or group owner and the joiners. The former uses a dedicated channel to send discovery messages, whereas the latter send their requests in the channel specified by the initiator. The study shows that defining two device categories and using a dedicated discovery channel improves the discovery time. The authors in [35] propose a design for the discovery message to compensate for link failures and reduce the control overhead.

Such message contains the identifier of the latestkdevices that have success- fully transmitted the discovery message. In [36], a proposal to dedicate a small portion of the resources to new devices appearing in the network is presented. By using this approach, the discovery message of the newcomers can be transmitted with a shorter delay, thus reducing their discovery time.

Given the fact that FD technology is viable, it can be used in D2D communication to overcome the HD limitation of not being able to listen to discovery messages from neighboring devices while transmitting the own message.

The strategy presented in [37] to reduce idle slots and collisions uses FD to detect the activity of other devices. The study assumes that a device stops transmitting the discovery message when it has been discovered. However, this assumption may not be valid when mobility or dynamic (de)activation of

(36)

devices is considered. In these kind of scenarios, transmitting the discovery message continuously is required. The authors in [38] combine FD with com- pressed sensing to overcome the drawbacks introduced by HD transmission mode and single packet reception. The presented results show that the discovery phase can be completed in a single time slot. However, the number of considered neighbors is rather limited and the design of the feedback mechanism for the transmission of acknowledgments is not addressed. Finally, [39]

evaluates FD with directional antennas. Each device selects a transmission direction randomly at each time slot. This mechanism is effective when two devices coincide in the same direction. However, the time scale is too long to meet the requirements defined for future systems [13], given the proposed time setting and the fact that only two devices can discover each other in each transmission.

It is important to note that, independently of the transmission mode (HD or FD), the conventional approach is to define a fixed transmission probabil- ity or periodicity to transmit the discovery message. This approach may be adequate in networks where the control latency is not a limitation. However, in the future 5G RAT, strict requirements are posed in terms of control and data plane latencies. Consequently, an appropriate design for the system to be able to meet the defined requirements is needed.

4 Scope and Objectives of the Thesis

As described in the previous sections, one of the main research areas in mobile communication is the design of a 5G system to accommodate the expected wireless mobile traffic demands by 2020. Since the standardization process is still ongoing and there is not a clear design for the future RAT, this dissertation is based on the 5G Nokia Bell Labs vision. In particular, the eMBB and the URLLC use cases are addressed.

The eMBB use case is studied assuming the 5G indoor small cell concept, optimized for dense local area deployments. The concept is presented in [16, 40, 41] and is described in detail in the chapter entitled5G Small Cell System Overview. It was originally proposed as a TDD system, because of its flexibility and the possibility of exploiting unpaired frequency bands, oper- ating at centimeter-wavefrequencies. Techniques to provide synchronization in time and frequency are considered, as well as MIMO transceivers and interference suppression receivers. A key element to optimally accommodate the traffic and improve the spectral efficiency is to exploit the adaptability of TDD to provide flexibility in the transmission direction slot assignment, i.e., dynamic TDD. However, this strategy brings challenges in terms of interference variability, which affects the signal reception. Another approach to

(37)

4. Scope and Objectives of the Thesis

achieve higher spectral efficiency is to use FD technology, given its promise of doubling the throughput of traditional HD systems. Nevertheless, the constraints that compromise the gain that FD can provide over dynamic TDD have to be addressed.

The theoretical potential of FD technology has been studied in terms of throughput. However, such technology could also bring large benefits in terms of latency. For this reason, FD is studied for the URLLC use case, with focus on D2D communication. D2D allows devices to communicate directly among them, thus being an attractive solution for offloading the infrastructure and accommodating the massive increase of mobile devices and the explosion of new services. In D2D, before the data exchange procedure, devices must discover their peers. A critical challenge in this type of communication is how to accelerate the discovery process in order to meet the strict latency target of 10 milliseconds defined for next generation access technologies [13].

To address the described challenges, this thesis presents detailed performance evaluation and Radio Resource Management (RRM) solutions, i.e., all the strategies to efficiently utilize the radio resources, such as scheduling, decision of the transmission parameters or resource allocation. More precisely, these solutions are designed for the proposed 5G small cell system to improve the capacity and accommodate the exponentially growing mobile data traffic, and for direct D2D communication to reduce the discovery time. Figure 1.7 shows the scope of the thesis.

Fig. 1.7:Scope of the thesis. Study cases highlighted in gray.

(38)

The hypothesis (H) and research questions (Q) that are addressed in this dissertation are listed below:

H1 Dynamic TDD [42] has the potential of achieving higher system throughput. However, providing flexibility in the transmission direction on a time slot basis leads to larger SINR variability. Since the system performance is affected by the remaining interference that advanced receivers cannot suppress, stabilization mechanisms are required to further enhance the system.

Q1 Can fast recovery and link adaptation mechanisms deal with packet losses caused by such increased interference variability?

H2 Traditional HD systems allow a node either to transmit or receive at a given time instant. By using FD technology, which allows simultaneous transmission and reception in the same frequency band, the system throughput may, theoretically, be doubled compared to HD systems.

However, in practice, there might be limitations in achieving the double throughput gain.

Q2 Which are these limitations and how do they impact the FD performance?

H3 Given the limited gains of FD in interference dominant scenarios, new applications where FD can bring larger benefits are studied. In direct D2D communication, devices must discover their neighbors prior to the actual data exchange. This procedure must be completed within the latency requirements, which is set to 10 milliseconds for 5G [13].

Q3 Is it possible to achieve such requirement? How can FD accelerate the device discovery procedure?

These research questions are respectively addressed in parts II, III and IV of the dissertation. Part II and III of the dissertation focus on investigating mechanisms to enhance the throughput and the delay of indoor ultra-dense small cell networks. In Part II, recovery mechanisms and link adaptation techniques have been studied to mitigate the drawback introduced by dynamic TDD, i.e., the increased interference variability [42]. Mechanisms such as HARQ and Outer Loop Link Adaptation (OLLA) are evaluated in the proposed 5G TDD system. In Part III, the emerging FD technology is considered as a potential candidate for indoor small cells. The performance of FD is studied considering dynamic TDD as the baseline system performance.

Two types of FD are studied: when both the Base Station (BS) and the UE can exploit simultaneous transmission and reception (bidirectional FD), and only when the BS is FD capable (BS FD). The constraints that prohibit to

(39)

5. Research Methodology

achieve the theoretical FD gain are investigated and analyzed, individually and jointly. Part IV of the dissertation studies how to provide fast discovery in the direct D2D communication. The discovery procedure can be performed autonomously by the devices or with the involvement of the infrastructure.

This work focuses on how to provide autonomous fast discovery for ad-hoc type of networks, where the infrastructure is not involved in the process.

Further details about the system design will be given in Part IV of the dissertation. The potential of FD technology in meeting the latency requirements defined in [13] is investigated.

5 Research Methodology

During the studies, a scientific methodology was employed to carry out the studies. Such methodology aims at dividing the research process into several phases:

1. Identify the research question to answer.

2. Literature review to get acquainted with the research area and related state of the art.

3. Design of solutions aiming at solving the identified problem.

4. Implementation and testing of the feature/mechanism through the simulation tool.

5. Collection of the simulation results and its corresponding dissemination in the form of a scientific article.

The target of this work is to provide a meaningful insight into how 5G would perform in 2020, assuming realistic conditions and a complete system, taking into account some of the envisioned technology components. Given the complexity of the work, it becomes unfeasible to perform the studies from a theoretical point of view. Consequently, the considered approach to evaluate the work of the thesis is through Monte Carlo system level simulations [43]. By using this methodology, meaningful results via comprehensive simulation campaigns can be extracted.

To evaluate the performance of the 5G system in ultra-dense small cells, a C++ event-driven simulator is used. It includes the implementation of the Open Systems Interconnection (OSI) protocol stack and the proposed 5G system design [41]. Figure 1.8a shows the layers which are implemented in the simulator. In the application layer, several traffic models are available, e.g.,

(40)

File Transfer Protocol (FTP) [44] and full buffer; in the transport layer, both User Datagram Protocol (UDP) and Transmission Control Protocol (TCP) are implemented; at the network level, the Acknowledged Mode (AM) and Un- acknowledged Mode (UM) modes defined for Radio Link Control (RLC) [45]

are included; in the Medium Access Control (MAC) layer, mechanism such as HARQ, scheduling and link and rank adaptation are available; finally, at the physical layer, two types of receivers are implemented, the Minimum Mean Square Error (MMSE)-Maximum Ratio Combining (MRC) receiver and the MMSE-Interference Rejection Combining (IRC) receiver. The modeling of the physical layer is out of the scope of this thesis, and both receiver models are extracted from [46]. SI cancellation is assumed ideal, and its implementation is not included in the simulator. The work carried out in this studies mainly focuses on schemes for the MAC layer, considering the interaction with the already existing transport and network layers. For more details regarding the simulator, please refer to the Appendices of [42].

The evaluation of the D2D study case is done using a Matlab simulator which uses some of the 5G concepts proposed in [41]. The focus of this study is on designing novel MAC protocols, as shown in Figure 1.8b. The simulator includes the ideal modeling of two receivers: abasicreceiver that treats interference as noise, and one that ideally suppress a certain amount of interfering streams, based on the degrees of freedom in the MIMO antenna domain. Fi- nally, the implementation of a SI cancellation model is not implemented since it is considered ideal.

(a)Ultra-dense small cells. (b)D2D communication.

Fig. 1.8:Layers from the OSI protocol stack where the study focus on.

(41)

6. Contributions and publications

6 Contributions and publications

The main contributions of this study can be summarized as follows:

1. Evaluating the potential of recovery mechanisms and link adaptation on the dynamic TDD 5G system.

Dynamic TDD allows to optimally accommodate the traffic, but also brings challenges in terms of interference variability, since the interfering source may change at each time slot. Although the use of interference suppression receivers has been proven to mitigate the drawback introduced by dynamic TDD, other mechanisms are required to cope with the unstable interference. HARQ and OLLA are introduced to the 5G system design to evaluate their potential in dealing with the unpredictable interference.

Results show that recovery mechanisms at lower layers are beneficial for the system. On the other hand, OLLA shows limited gain but it does not come at the expenses of increased overhead or delay.

2. Introducing a RRM module to provide support for FD technology into the envisioned 5G system.

Given the recent advances in the transceiver design, FD communication is positioned as a potential technology component for 5G systems. There are three different types of FD communication: when both the BS and the UE are FD capable (bidirectional FD); when only the BS is able to exploit such technology for data transmission (BS FD); and when only the BS is FD capable and uses the technology for relaying user data (relay FD). The module introduced allows for the evaluation of the first two FD types. It is composed by two submodules. Firstly, the optimal transmission direction is extracted per each node. Secondly, the decision of which user or pair of users is going to be scheduled is obtained. By separating functionali- ties, the number of operations to perform can be reduced and hence the processing time and the system complexity.

3. Identifying the constraints that limit the achievable gain of FD in 5G small cell networks.

The performance of bidirectional FD and BS FD is evaluated in the 5G ultra-dense small cell system, considering dynamic TDD as the baseline system performance. An exhaustive analysis is carried out, identifying the constraints that affect the gain that FD can provide over dynamic TDD.

In addition, an evaluation of the interaction of FD with higher protocols, such as TCP, is provided. The overall analysis is carried out assuming

(42)

ideal self-interference cancellation, thus providing an upper bound of the achievable FD gain in indoor small cell networks.

The evaluation of the potential of FD in 5G indoor small cell networks shows that this technology brings limited gains over dynamic TDD. Conse- quently, identifying other applications where FD could provide larger im- provements is required.

4. Evaluating the potential of FD to accelerate the discovery procedure in D2D communication.

Direct D2D communication is becoming popular given its potential to offload traffic from the infrastructure and to allow for a faster communication among devices. However, prior to the exchange of data, devices must discover their peers. This procedure, known as device discovery, should be completed in 10 milliseconds to meet the requirements imposed by 3GPP [13]. The potential of FD technology to speed up such process and to satisfy the latency requirements is evaluated. The analysis shows that the use of interference cancellation receivers is a requirement to achieve low latency, and in that case, FD technology is able to further reduce the discovery time.

5. Designing a protocol supporting FD technology to provide fast discov- ery in D2D communication.

Given the potential of FD in accelerating the discovery phase in D2D communication, a protocol to perform such procedure is proposed. A technique to estimate the number of neighboring devices alongside a signaling exchange mechanism to reduce the network interference are introduced.

Furthermore, several approaches on how to use the signaled information are proposed and evaluated. The study compares the performance of HD transmission mode with FD communication, proving that FD technology achieves lower latency, achieving the 3GPP requirements in most of the evaluated scenarios.

In addition to these contributions, exhaustive development of the sim- ulators was performed during the entire period of the studies. The C++

event-driven simulator used to extract the results for Part II and III of the dissertation was originally implementing the LTE system, and it was modified to include the envisioned 5G design [41]. The main contributions to the simulator development are listed below:

(43)

6. Contributions and publications

• 5G processing delay

The control and data transmission timing, processing delay and the proposed frame structure defined in the concept [41] were implemented in the simulator to extract realistic results.

• HARQ and link adaptation

An asynchronous and non-adaptive implementation of the HARQ mechanism was implemented in the simulator. In terms of link adaptation, an scheme to obtain the Modulation and Coding Scheme (MCS) for transmission based on the average of the latestkSINR samples in logarithmic scale was implemented, as well as a static and a dynamic OLLA mechanism.

• Full duplex transmission

The procedure for the signal transmission and reception in the simulator was modified to allow for both HD and FD transmission modes. The modification was included in both control and data parts, although in this dissertation such functionality is only exploited in the data part.

• Multi-user functionality

The limitation of a single user per small cell did not allow for evaluating the FD case where the AP is the only node which can exploit simultaneous transmission and reception. The multi-user functionality affected the entire simulator, e.g., the generation of data flows or the RLC data aggregation and acknowledgment transmission.

• User and transmission direction scheduler

Introducing the multi-user functionality required a mechanism to decide the transmission mode, the transmission direction in case of HD and the scheduled user(s).

• Ideal RLC control transmission

The RLC acknowledgment transmission occupied a single Transmission Time Interval (TTI) and the whole bandwidth. Since the generated overhead was unrealistic, an ideal RLC acknowledgment transmission through the control channel was implemented.

The post-processing of the data obtained from the simulator was performed with Matlab to extract and present the results. Finally, Part IV of the thesis is carried out with an own Matlab simulator, developed entirely and only for the purpose of the last study in D2D communication.

The following publications were authored or co-authored in relation with this study:

(44)

Part I:

• Improving Link Robustness in 5G Ultra-Dense Small Cells by Hybrid ARQ.

Gatnau, Marta; Catania, Davide; Frederiksen, Frank; Cattoni, Andrea Fabio;

Berardinelli, Gilberto; Mogensen, Preben. IEEE 11th International Sympo- sium on Wireless Communications Systems (ISWCS), 2014.

• Dynamic Outer Loop Link Adaptation for the 5G Centimeter-Wave Con- cept. Gatnau, Marta; Catania, Davide; Frederiksen, Frank; Cattoni, Andrea Fabio; Berardinelli, Gilberto; Mogensen, Preben. IEEE 21th European Wire- less Conference, 2015.

• The Potential of Flexible UL/DL Slot Assignment in 5G Systems. Cata- nia, Davide; Gatnau, Marta; Cattoni, Andrea Fabio; Frederiksen, Frank;

Berardinelli, Gilberto; Mogensen, Preben. IEEE 80th Vehicular Technology Conference (VTC) Fall, 2014.

• Flexible UL/DL in Small Cell TDD Systems : A Performance Study with TCP Traffic. Catania, Davide;Gatnau, Marta; Cattoni, Andrea Fabio; Fred- eriksen, Frank; Berardinelli, Gilberto; Mogensen, Preben. IEEE 81st Vehic- ular Technology Conference (VTC) Spring, 2015.

Part II:

• Full Duplex Communication Under Traffic Constraints for 5G Small Cells.

Gatnau, Marta; Catania, Davide; Berardinelli, Gilberto; Mahmood, Nurul Huda; Mogensen, Preben. IEEE 82nd Vehicular Technology Conference (VTC) Fall, 2015.

• Can Full Duplex Boost Throughput and Delay of 5G Ultra-Dense Small Cell Networks?. Gatnau, Marta; Berardinelli, Gilberto; Mahmood, Nurul Huda; Mogensen, Preben. IEEE 83rd Vehicular Technology Conference (VTC) Spring, 2016.

• Impact of Transport Control Protocol on Full Duplex Performance in 5G Networks. Gatnau, Marta; Berardinelli, Gilberto; Mahmood, Nurul Huda;

Mogensen, Preben. IEEE 83rd Vehicular Technology Conference (VTC) Spring, 2016.

• On the Potential of Full Duplex Performance in 5G Ultra-Dense Small Cell Networks. Gatnau, Marta; Fleischer, Marko; Berardinelli, Gilberto; Mah- mood, Nurul Huda; Mogensen, Preben; Heinz, Helmut. IEEE European Signal Processing Conference (EUSIPCO), 2016.

• Analyzing the Potential of Full Duplex in 5G Ultra-Dense Small Cell Net- works. Gatnau, Marta; Berardinelli, Gilberto; Mahmood, Nurul Huda;

(45)

7. Thesis Outline

Fleischer, Marko; Mogensen, Preben; Heinz, Helmut. EURASIP Journal on Wireless Communications and Networking - Special issue: Full-Duplex Radio: Theory, Design, and Applications, 2016. Status:submitted.

• Evaluating Full Duplex Potential in Dense Small Cells from Channel Mea- surements. Berardinelli, Gilberto; Assefa, Dereje; Mahmood, Nurul Huda;

Gatnau, Marta; Sørensen, Troels Bundgaard; Mogensen, Preben Elgaard.

IEEE Vehicular Technology Conference (VTC) Spring, 2016.

Part III:

• Can Full Duplex reduce the discovery time in D2D Communication?. Gat- nau, Marta; Berardinelli, Gilberto; Mahmood, Nurul Huda; Soret, Beatriz;

Mogensen, Preben. IEEE 13th International Symposium on Wireless Com- munications Systems (ISWCS), 2016.

• Providing Fast Discovery in D2D Communication with Full Duplex Tech- nology. Gatnau, Marta; Berardinelli, Gilberto; Mahmood, Nurul Huda;

Soret, Beatriz; Mogensen, Preben. Springer 9th International Workshop on Multiple Access Communications (MACOM), 2016. Status:submitted.

• Radio resource management for V2V discovery. Soret, Beatriz; Gatnau, Marta; Kovács, István Z.; Martín-Vega, Francisco Javier; Berardinelli, Gilberto;

Mahmood, Nurul Huda. IEEE Vehicular Technology Magazine, 2016. Sta- tus:submitted.

7 Thesis Outline

This Ph.D. dissertation is composed as a collection of papers. For this reason, the contributions and outcomes extracted from this work are presented in the form of accepted or submitted conference and journal articles, included in Parts II, III and IV of the thesis. Each of these parts consist of a general overview, i.e., the problem definition, the assumptions and the main find- ings. Part I of the dissertation provides an introduction to the study and a description of assumed the 5G system design. Part V exposes the main conclusions and future research work. Finally, Part VI corresponds to the Appendix, which includes the conference articles not included in the main body of the dissertation. The outline of the thesis is as follows:

Part I. Includes the current chapter and a detailed description of the proposed 5G indoor small cell system.

(46)

References

Part II. Analyses the benefits of using recovery mechanisms and link adaptation schemes to mitigate the drawbacks introduced by dynamic TDD. This part is composed by papers A and B.

Part III. Studies the potential of FD technology in improving the 5G system performance, in terms of throughput and delay. For the purpose of the reader, only paper C is included in this part, since it provides a complete overview of the four accepted conference papers included in Part VI Appendix.

Part IV. Examines the requirements to provide fast discovery in direct D2D communication and achieve the strict latency target requirements. Papers D and E compose this part.

Part V. Concludes the Ph.D. dissertation and provides recommenda- tions for future research work.

Part VI. Includes additional first-author papers complementing the dissertation.

A number of abbreviations are used on this thesis, spelled out in their first appearance. We recommend the reader to use the List of Abbreviations included after the Table of Contents while reading the thesis. A reference list is included at the end of each chapter and paper. Note that the references which are cited within each chapter are not necessarily represented by the same number in all chapters.

References

[1] Cisco, “White paper: Cisco visual networking index: Global mobile data traffic forecast update, 2015-2020,” Feb 2016.

[2] J. G. Andrews, S. Singh, Q. Ye, X. Lin, and H. Dhillon, “An overview of load bal- ancing in hetnets: old myths and open problems,”IEEE Wireless Communications, vol. 21, no. 2, pp. 18–25, Apr 2014.

[3] A. Ghosh et al., “Millimeter-wave enhanced local area systems: A high-data- rate approach for future wireless networks,” IEEE Journal on Selected Areas in Communications, vol. 32, no. 6, pp. 1152–1163, Jun 2014.

[4] S. Larew, T. A. Thomas, M. Cudak, and A. Ghosh, “Air interface design and ray tracing study for 5G millimeter wave communications,” in IEEE Globecom Workshops (GC Wkshps), Dec 2013, pp. 117–122.

Aalborg Universitet Radio Resource Management for 5G Small Cells in Unpaired Spectrum With Focus on Dynamic TDD and Full Duplex Technology Gatnau, Marta